Towards Healing the Blindness of Score Matching
Abstract
Score-based divergences have been widely used in machine learning and statistics applications. Despite their empirical success, a blindness problem has been observed when using these for multi-modal distributions. In this work, we discuss the blindness problem and propose a new family of divergences that can mitigate the blindness problem. We illustrate our proposed divergence in the context of density estimation and report improved performance compared to traditional approaches.
Cite
Text
Zhang et al. "Towards Healing the Blindness of Score Matching." NeurIPS 2022 Workshops: SBM, 2022.Markdown
[Zhang et al. "Towards Healing the Blindness of Score Matching." NeurIPS 2022 Workshops: SBM, 2022.](https://mlanthology.org/neuripsw/2022/zhang2022neuripsw-healing/)BibTeX
@inproceedings{zhang2022neuripsw-healing,
title = {{Towards Healing the Blindness of Score Matching}},
author = {Zhang, Mingtian and Key, Oscar and Hayes, Peter and Barber, David and Paige, Brooks and Briol, Francois-Xavier},
booktitle = {NeurIPS 2022 Workshops: SBM},
year = {2022},
url = {https://mlanthology.org/neuripsw/2022/zhang2022neuripsw-healing/}
}